Hung Huynh, 2 Pierce K. H. Chow, London L. P. Ooi, and Khee-Chee Soo

Transcription

1 Vol. 13, , March 2002 Cell Growth & Differentiation 115 A Possible Role for Insulin-like Growth Factor-binding Protein-3 Autocrine/Paracrine Loops in Controlling Hepatocellular Carcinoma Cell Proliferation 1 Hung Huynh, 2 Pierce K. H. Chow, London L. P. Ooi, and Khee-Chee Soo Laboratory of Molecular Endocrinology [H. H.] and the Department of Surgical Oncology [P. K. H. C., L. L. P. O., K-C. S.], National Cancer Centre, Singapore and Department of General Surgery, Singapore General Hospital [P. K. H. C., L. L. P. O., K-C. S.], Singapore Abstract Hepatocellular carcinoma (HCC) is a common malignancy, but treatment outcomes have generally remained poor. Specific factors important for the pathogenesis of HCC are incompletely understood. Insulin-like growth factors (IGFs) are potent autocrine and paracrine mitogens for liver cancer cell proliferation, and their bioactivity is reduced by IGFbinding protein 3 (IGFBP-3). In the present study, we report that IGFBP-3 protein levels were either undetectable (28.5%) or low (71.5%) in human HCC samples examined compared with matched nonneoplastic liver tissue by Western blotting. IGFBP-3 was localized to nontumor liver cells by immunohistochemistry with greater immunointensity than neoplastic liver cells. Levels of type I receptor (IGF-IR) were found to be low in 39% of human HCC samples examined compared with matched nontumor tissues. IGF-II was overexpressed in 32%, whereas IGF-I expression was decreased in 100% of HCC samples. In vitro studies revealed that IGF-I and IGF-II induced HepG2 cell proliferation in a dose-dependent manner. Treatment of HepG2 cells with either human recombinant IGFBP-3 (hrigfbp-3) or IGF-II antibody led to a significant reduction in cell proliferation. Cotreating these cells with hrigfbp-3 significantly attenuated the mitogenic activity of IGF-I. IGF-Iinduced phosphorylation of IGF-IR subunit, IRS-1, mitogen-activated protein kinase, Elk-1, and Akt-1 as well as phosphatidylinositol 3 -kinase activity was significantly attenuated when hepg2 cells were pretreated with hrigfbp-3. Our data indicate that loss Received 11/1/01; revised 12/27/01; accepted 1/7/02. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by grants from National Medical Research Council of Singapore (NMRC/0541/2001) and SingHealth Cluster Research Fund (EX 008/ 2001; to H. H.). 2 To whom requests for reprints should be addressed, at Molecular Endocrinology Laboratory, National Cancer Centre of Singapore, 11, Hospital Drive, Singapore Phone: (65) ; Fax: (65) ; cmrhth@nccs.com.sg. of autocrine/paracrine IGFBP-3 loops may lead to HCC tumor growth and suggest that modulating production of the IGFs, IGFBP-3, and IGF-IR may represent a novel approach in the treatment of HCC. Introduction HCC 3 is one of the most common malignancies worldwide, but treatment outcomes for HCC have remained generally poor. The majority of patients with HCC have inoperable disease with very poor prognosis (1). Survival in patients with curative resection carried out at dedicated centers is between 35 50% at 5 years and much lower elsewhere (2, 3). Long-term survival is uncommon because of the frequent presence of recurrence, metastasis, or the development of new primaries (4, 5). There is also currently no accepted adjuvant or palliative treatment modalities that have been conclusively shown to prolong survival in HCC (6). Recent research on IGF-I and IGF-II have shown these to be potent mitogens for human hepatoma cells (7), and both IGF-I mrna (8) and IGF-IRS (7, 8) have been detected in human hepatoma cell lines. In vitro studies on human HCC cell lines HuH-7 and HepG2 have demonstrated that these cells secrete IGF-II and that the inhibition of IGF-II expression in these cells led to a reduction in cell proliferation (8). IGF-I and IGF-II mrna have also been detected in vivo in HCC, but the levels compared with matched nontumorous adjacent hepatic tissue were lower than in HCC for IGF-I mrna (9) and higher in HCC for IGF-II mrna (9 11). Furthermore, IGF-II fetal transcripts (5.6 and 4.5 kb) were found in HCC (9), whereas normal IGF-II transcripts were regressed (10). Coexpression of IGF-II protein and Ki-67 antigen in tumor cells were observed suggesting that IGF-II acts as an autocrine or paracrine growth factors in these tumors (10). Overexpression of IGF-II in transgenic mice has been shown to increase serum IGF-II and HCC (12). These observations suggest an important role for IGF-II in hepatocarcinogenesis. Both IGF-I and IGF-II bind with high affinity to specific IGFBPs, which modulate their bioactivity. At least six IGFBPs have been described (reviewed in Refs. 13, 14). Expression of genes encoding the various IGFBPs has been observed in many tissues and is subject to intricate physiological regulation (reviewed in Refs. 13, 14). IGFBP-3 is the most abun- 3 The abbreviations used are: HCC, hepatocellular carcinoma; IGF, insulin-like growth factor; IGF-IR, IGF-I receptor; IGFBP, insulin-like growth factor-binding protein; hrigfbp-3, human recombinant IGFBP-3; PI3k, phosphatidylinositol 3 -kinase; MAPK, mitogen-activated protein kinase; IRS-1, insulin receptor substrate 1; Akt, serine/threonine protein kinase; GH, growth hormone; PSF, phenol-red serum-free; SFM, serum-free medium.

2 116 IGFBP-3 in Hepatocellular Carcinoma Fig. 1. IGFBP-3 expression in adjacent normal liver tissue and HCC tumors. Tissue lysates from normal adjacent liver tissue and HCC tumors were analyzed by Western blotting. Blots were incubated with antihuman IGFBP-3 (B, D, and F) and -tubulin (A, C, and E) antibodies. Representative samples are shown. All normal adjacent liver tissues (N) had high levels of IGFBP-3, whereas IGFBP-3 protein was either underexpressed or absent from HCC tumors (T). Serum IGFBP-3 served as a positive control. Table 1 Expression of IGFs, IGFBP-3, and IGF-IR in normal adjacent liver tissues and HCC tumors Genes examined Normal adjacent liver tissues (n 28) HCC tumors (n 28) Intense staining of IGF-I (100%) 18/28 (64%) 10/28 (36%) Intense staining of IGF-II (100%) 19/28 (68%) 9/28 (32%) Expression levels of IGFBP-3 levels: (compared with normal and adjacent tissue of the same patient) Normal 28/28 (100%) 0/28 (0%) Low 0/28 (0%) 8/28 (28.5%) Undetectable 0/28 (0%) 20/28 (71.5%) IGF-IR expression levels: (compared with normal and adjacent tissue of the same patient) Normal 28/28 (100%) 17/28 (61%) Low 0/28 (0%) 11/28 (39%) dant IGFBP in the circulation, where it forms a M r 150,000 complex with an acid-labile subunit and IGF-I or IGF-II (14). The IGFBP-3 gene is expressed in many tissues, and IGFBP-3 has affinities for IGFs that are either equal to or stronger than those of the IGF receptors and, therefore, inhibit the IGFs by sequestration in the extracellular compartment (reviewed in Refs. 13, 14). Recent evidence also demonstrates that IGFBP-3 has growth-inhibitory activity that is independent of its IGF binding properties (reviewed in Refs. 14). Thus, existing evidence suggests that a better understanding of IGFBP-3 may allow modulation of its activities and the potential for therapeutic applications in the control of HCC. Results In the adult human, the liver is the main source of circulating IGF-I and IGFBP-3 (13, 14). Because local expression of IGF-I is important in autocrine and paracrine stimulation, and the IGFBP-3 appears to decrease the mitogenic activity of free IGF-I, presumably by competing with type I IGF receptors for the ligand, the IGFBP-3 levels in the HCC samples were determined. Total tissue lysates from both HCC cancers and adjacent morphologically normal liver cells were subjected to Western blot analysis. The resulting blots were incubated with IGFBP-3 antibody. Fig. 1 shows that IGFBP-3 protein (a complex spanning M r 38,000 42,000) was either absent or low in HCC compared with adjacent normal tissues. IGFBP-3 bands were undetectable in 28.5% (8 of 28) of HCC samples examined (Table 1). Although the remaining tumors (20 of 28; 71.5%) expressed IGFBP-3, the level was significantly lower than that observed in normal adjacent tissue (P 0.01). Densitometric scanning showed IGFBP-3 levels to be fold higher in normal liver tissue than in HCC. Subsequent blotting with anti- tubulin antibody showed relatively equal amounts of total protein loaded per lane (Fig. 1).

3 Cell Growth & Differentiation 117 Fig. 3. Immunostaining of normal liver and HCC tumors for IGF-I and IGF-II. Normal adjacent liver tissue (A and C) and HCC tumors (B and D) were stained with antihuman IGF-I (A and B) or antihuman IGF-II (C and D) as described in Material and Methods. Representative samples are shown. Normal adjacent liver tissue shows intense staining for IGF-I, whereas low staining signals are seen for HCC cells. HCC tumors show more intense staining for IGF-II than normal adjacent tissue. (original magnification 400). Fig. 2. Immunostaining of normal liver (A) and HCC tumors (B) for IGFBP-3. Normal adjacent liver tissue and HCC tumors were stained with antihuman IGFBP-3 as described in Material and Methods. Adjacent normal liver tissue, showing intense expression of normal cells for IGFBP-3, whereas very low staining signal was observed for the morphologically disorganized HCC cells. Representative staining are shown. (original magnification 400). To confirm the above observation, immunohistochemical staining of HCC and normal adjacent liver tissues was performed using antihuman IGFBP-3 antibody, which recognized both intact and fragmented IGFBP-3 protein. IGFBP-3 protein was clearly localized to almost all of the cells in the adjacent normal tissue (Fig. 2A). However, this signal was absent (Fig. 2B) in HCC, indicating that IGFBP-3 expression was either inactivated or decreased in these neoplastic cells. To determine whether HCC also expressed IGF-I and IGF- II, HCC and adjacent normal liver tissues were stained with anti-igf-i and anti-igf-ii antibodies, respectively. Although normal adjacent tissues stained positively for IGF-I (Fig. 3A), the expression was significantly decreased in HCC (Fig. 3B). Approximately 64% (18 of 28) and 36% (10 of 28) of HCC samples examined had moderate and low IGF-I expression, respectively (Table 1), which is consistent with a previous report (9). IGF-II staining (Fig. 3, C and D), in contrast, was approximately equal (68%; 19 of 28) or more intense (32%; 9 of 28) in HCC compared with normal liver tissue (Table 1). These observations are in agreement with previous reports demonstrating overexpression of IGF-II in HCC (9 11). Our data suggests that IGF autocrine and paracrine loops exist in HCC. Because IGFs stimulate growth responses in liver cells by binding to the IGF-IR, IGF-IR levels in HCC and normal adjacent tissue samples were examined. Approximately 39% of HCC tumors had slightly lower IGF-IR levels than nonneoplastic adjacent liver tissues (Fig. 4; Table 1). When normalized for -tubulin, IGF-IR expression was not significantly different between normal tissues and HCC (P 0.01). To test the hypothesis that IGF-I and IGF-II play a role in mediating tumor cell growth, HepG2 cells were treated with various concentrations of human recombinant IGF-I or IGF-II for 48 h. Fig. 5A shows dose-dependent growth stimulation by IGF-I. Similar effects were observed when HepG2 cells were treated with IGF-II (data not shown). Four-fold increase in cell number was observed at the dose of 25 ng/ml IGF-I (P 0.01). Because IGFBP-3 protein was undetectable or lost in the majority of HCC samples (Fig. 1) and IGFBP-3 has been shown to inhibit cancer cells in an IGF-independent pathway (14), we attempted to demonstrate the antiproliferative action of IGFBP-3 on HepG2 cells. Approximately 40% and 52% inhibition of basal proliferation was obtained via 250 ng/ml and 500 ng/ml of hrigfbp-3, respectively (P 0.01; Fig. 5B). Concentrations of hrigfbp-3 up to 1 g/ml did not inhibit additional proliferation. To determine whether the proliferative action of IGF-I can be attenuated in the presence of IGFBP-3, HepG2 cells were treated with hrigfbp-3 in the presence and absence of IGF-I for 48 h. As shown in Fig. 5C, treatment of HepG2 with 25 ng/ml IGF-I for 48 h resulted in a 2.8-fold increase in cell number (P 0.01). IGF-induced HepG2 proliferation was significantly attenuated (P 0.01) in the presence of 250 ng/ml IGFBP-3 and completely abolished at the concentra-

4 118 IGFBP-3 in Hepatocellular Carcinoma Fig. 4. IGF-IR expression in adjacent normal liver tissue and HCC tumors. Tissue lysates from normal adjacent liver tissue (N) and HCC tumors (T) were analyzed by Western blotting as described in Materials and Methods. Blots were incubated with anti-igf-ir (B and D) and -tubulin (A and C) antibodies. Note that in few pairs of samples, normal adjacent liver tissue had nonsignificantly higher IGF-IR than HCC tumors when normalized for tubulin. Fig. 5. Effects of IGFs, IGFBP-3, and IGF-II antibody on HepG2 cell proliferation. Cells were cultured as described in Materials and Methods. Cells were incubated with serum-free MEM, and indicated reagents of IGF-I for 48 h. Cell number was determined as described (15). Means of triplicate experiments were plotted. Bars with different letters are significantly different from one another at (P 0.01, Mann-Whitney U test). SE at each point was 15%; bars, SE. A, effect of IGF-I on HepG2 proliferation. IGF-I significantly increased HepG2 cell number. B, effect of IGFBP-3 on HepG2 cell proliferation. Growth of HepG2 cells was significantly inhibited by exogenous IGFBP-3. C, effect of hrigfbp-3 on IGF-I-induced HepG2 cell proliferation. Cells were incubated with either SFM or 25 ng/ml IGF-I in the presence of various concentrations of hrigfbp-3 for 48 h. IGF-I-induced HepG2 proliferation was significantly attenuated by hrigfbp-3. D, effect of IGF-II and IGF-II antibody on HepG2 cell proliferation. Cells were incubated with SFM, 25 ng/ml human recombinant IGF-II, rabbit preimmune serum (1:800 dilution), rabbit antihuman IGF-II antibody (2 g/ml), and 25 ng/ml IGF-I in combination with rabbit antihuman IGF-II (2 g/ml). IGF-II-induced proliferation rates were significant decreased in the presence of anti-igf-ii antibody. Note that IGF-II antibody also inhibits HepG2 cell growth. Bars with different letters are significantly different from one another at (P 0.01). Experiments were repeated three times with similar results. Results shown are representative of two independent experiments; bars, SE. tion of 500 ng/ml. This result suggests that IGFBP-3 attenuated IGF-I-induced HepG2 proliferation by reducing IGF-I bioavailability. To test the hypothesis that autocrine production of IGF-II by liver cancer cells plays a role in mediating liver cancer cell growth, HepG2 cells, which has been shown to secrete IGF-II (8), were treated with anti-igf-ii, IGF-II, or with both combined for 48 h. Fig. 5D shows a 3-fold induction of basal proliferation by 25 ng/ml IGF-II (P 0.01). This induction was significantly attenuated by IGF-II antibody (P 0.01), whereas preimmune serum was not significantly affected. IGF-II antibody alone caused a 35% reduction in cell number. These results suggest that the rapid proliferation of liver cancer cells in vivo and in vitro may at least, in part, be a consequence of autocrine stimulation mediated by IGF-II expression. Because MAPK and PI3k are important for the effect of IGFs on growth and apoptosis, the mechanisms involved

5 Cell Growth & Differentiation 119 Fig. 6. Effects of IGF-I and hrigfbp-3 on the expression and phosphorylation of IGF-IR subunit, IRS-1, and MAPK in HepG2 cells. HepG2 cells were stimulated with IGF-I (25 ng/ml), hrigfbp-3 for 24 h, or pretreated with 250 ng/ml hrigfbp-3 for 24 h, washed, and stimulated with 25 ng/ml IGF-I. Lysates from these cells were immunoprecipitated with either anti- IGF-IR (B and C) or IRS-1 (D and E) antibodies as described in Materials and Methods. After SDS-PAGE, blots were immunoblotted with antibodies to IGF-IR (B), IRS-1 (D), and phosphotyrosine clone 4G10 (C and E). To detect total MAPK, phospho MAPK, and phospho Elk-1, cell lysates were subject to Western blotting as described in Materials and Methods. Blots were blotted with anti- tubulin (A), anti-mapk (F), phospho p44/42 MAPK (Thr202/Tyr204; G), and phospho Elk-1 (Ser383; H) antibodies. The figure is representative of three independent experiments. in the attenuation of IGF-I by IGFBP-3 on proliferation of human HepG2 cancer cells were investigated. HepG2 cells were preincubated with 250 ng/ml hrigfbp-3 for 24 h. The medium was removed, and cells were rinsed with serumfree medium and exposed to IGF-I for 10 min. Various components of IGF-IR signaling (IGF-IR, IRS-1, Akt, MAPK, and PI3k) were analyzed. Treatment with IGF-I resulted in a significant increase in tyrosine phosphorylation of IGF-IR subunit, IRS-1, and MAPK (P 0.01; Fig. 6, C, E, and G). Activation of MAPK by IGF-I also resulted in the activation of Elk-1 transcription factor (Fig. 6, H). Preincubation of HepG2 cells with hrigfbp-3 lead to a significant reduction in IGF-I-induced phosphorylation of IGF-IR, IRS-1, Elk-1, and MAPK (P 0.01). Treating HepG2 cells with IGF-I also resulted in an increase in the association of p85 with IRS-1 (data not shown). A 2.5-fold increase in basal PI3k activity by IGF-I was observed (P 0.01; Fig. 7A). IGFBP-3 alone also significantly reduced PI3k activity (P 0.01). Phosphorylation of Akt at Ser 473 was also induced by IGF-I (P 0.01; Fig. 7C). IGF-induced PI3k activity and Akt-1 phosphorylation were significantly attenuated by preincubating HepG2 cells with IGFBP-3 (P 0.01; Fig. 7C). Fig. 7. Effects of IGF-I and hrigfbp-3 on the expression and activity of PI3k and the phosphorylation of Akt-1 in HepG2 cells. HepG2 cells were stimulated with IGF-I (25 ng/ml) for 10 min, hrigfbp-3 for 24 h, or pretreated with 250 ng/ml hrigfbp-3 for 24 h, washed, and stimulated with 25 ng/ml IGF-I for 10 min. Lysates from these cells were immunoprecipitated with IRS-1 antibody as described in Materials and Methods. Immunocomplexes were used to determine PI3k activity associated with IRS-1 (A). Quantitative analysis of 32 P-labeled inositol 1,3,4 5 phosphate levels was determined by scanning of the blots densitometrically (B). Bars with different letters are significantly different from one another at (P 0.01); bars, SE. C, cell lysates were also submitted to SDS-PAGE, and the immunoblot was performed with anti- tubulin, anti-akt, which recognizes both phospho and dephospho-akt-1, and phospho-akt Ser (473) antibodies. Results shown are representative of two independent experiments. Discussion Very little information is available on the expression of IGFBP-3 in HCC. In the present study, we show that IGFBP-3 protein was lost or underexpressed in the majority of HCC. IGFBP-3 was intensely expressed in corresponding paired samples of normal liver cells. Because Kupffer cells are the main source of IGFBP-3 production in human liver (15), it is likely that these IGFBP-3-expressing cells are Kupffer cells. Because no low molecular weight IGFBP-3 bands were detected in Western blot analysis (data not shown), it is unlikely that proteolysis of IGFBP-3 is the cause of absent IGFBP-3 expression in HCC. The absence of IGFBP-3 in HCC was reinforced by immunohistochemical analysis showing that the normal hepatic cells expressed high levels of IGFBP-3 protein whereas HCC did not. The ability to detect low amount of IGFBP-3 in certain cancerous tissues by Western

6 120 IGFBP-3 in Hepatocellular Carcinoma blot analysis but not by immunohistochemistry could be because of blood contamination. All of the HCC samples significantly expressed less IGF-I than control liver. Unlike IGF-I, most of HCC samples retained IGF-II expression, and 32% of HCC samples significantly overexpressed IGF-II. IGF-IR levels were not significantly low in 39% of HCC samples. These observations suggest an autocrine/paracrine role of IGFs and IGFBP-3 in the proliferation of liver cancer cells in vivo. This suggestion was reinforced by in vitro experiments involving the IGFBP-3 and IGF antibodies. Blocking IGF-II autocrine function by treating HepG2 cells with IGF-II antibody significantly reduced cell-proliferation indicating that functional autocrine IGF-II loop exists in these cells. Treatment of HepG2 liver cancer cells with IGFBP-3 also led to growth inhibition, whereas treatment of HepG2 with IGF-I or IGF-II led to increased DNA synthesis. The growth-promoting effect of IGF-I and IGF-I-induced phosphorylation of key proteins in PI3k and MAPK pathways were significantly attenuated when HepG2 cells were pretreated with hrigfbp-3. Our results provide evidence for a paracrine/endocrine mechanism(s) by which IGFs and IGFBP-3 influence hepatoma cell proliferation. Because endogenous IGF-I and IGF-II are produced by both normal liver cells and HCC cells, the decrease or loss of IGFBP-3 production by HCC would allow more free IGFs to act in an autocrine or paracrine fashion to enhance tumor cell growth. Although expressions of other IGF-binding proteins in HCC such as IGFBP-1 and IGFBP-2 (data not shown) were also investigated, only IGFBP-3 levels were associated with HCC growth suggesting the possible involvement of IGFBP-3 in controlling HCC cell proliferation. The observations that HCC tumors express less IGF-I than control liver, whereas IGF-II expression is higher in a high proportion of HCC tumors, are consistent with previous reports that showed IGF-I mrna levels were lower in HCC as compared with nontumorous hepatic tissue from an adjacent area (9), whereas IGF-II mrna was higher tumor compared with normal liver (13, 14, 16). At the moment, the molecular mechanisms responsible for reduction in IGF-I and reactivation of IGF-II in HCC remain to be determined. It has been reported that hepatic IGF-I and IGF-IR expression is regulated by GH (13, 14) and estrogen (16). It is possible that the reduction in IGF-I and IGF-IR expression in HCC observed in the present study may be because of the loss of receptors for GH or estrogens. Experiments are under way to investigate this possibility. Currently, the molecular mechanisms by which the IGFBP-3 expression is lost or reduced in HCC are not known. Several reports have suggested that IGFBP-3 expression is regulated by GH (reviewed in Refs. 13, 14), antiestrogen (17), transforming growth factor (18), and retinoic acid (19). In HCC, receptors of these compounds may be lost or inactivated, leading to IGFBP-3 gene inactivation. It is possible that the absence of IGFBP-3 expression in HCC is a consequence of genetic alterations such as deletion, mutation, or inappropriate hypermethylation as described in a number of tumor suppressor genes (20, 21) and cyclin-dependent kinase inhibitor genes (22, 23), or chronic hepatitis and cirrhosis (24, 25). We are currently investigating the mechanisms responsible for silencing the IGFBP-3 gene and the relationship between hepatitis and cirrhosis with the loss of IGFBP-3 expression in HCC. In our experimental system, we also observed that IGFBP-3 inhibited proliferation of HepG2 cells and could be an important factor in a negative control system regulating human cancer cell growth in vivo. Because the gene encoding IGF-II is expressed in HepG2 (8), the observed inhibitory action of IGFBP-3 likely involves the reduction of bioavailability of endogenous IGF-II for cell surface receptor binding. It is also possible that IGFBP-3 inhibits HepG2 by direct growth inhibitory signal transduction pathways (26, 27). However, these mechanisms are not mutually exclusive and both may be relevant in vivo. Human liver cancer cell growth is inhibited by exogenous IGFBP-3. In addition, IGFBP-3 is a potent IGF-I antagonist. These properties suggest a possible therapeutic application for IGFBP-3 in the control of liver cancer cell proliferation. Also, the presence of IGFBP-3 in HCC specimens may possibly be a better marker for tumor stage, diagnosis, and prognosis than currently used histological parameters. Experiments are under way to analyze the association between levels of serum IGFBP-3 and tumor stage, metastasis, and other markers such as serum -feto protein in patients with HCC. Although many treatment modalities have been attempted in HCC, there is currently no proven practical therapy besides surgery, and attempts at both adjuvant therapy and therapy for inoperable HCC remains experimental (6). Because HCC is capable of endogenous production of both IGF-I and IGF-II (8 11), and hepatoma cells are very responsive to stimulation by these growth factors, agents that are capable of modulating endocrine, paracrine, and autocrine production of the IGFs and their receptors are likely to be useful in treatment of HCC. The present findings offer both a potential diagnostic parameter and as well as a potential novel strategy for HCC cancer endocrine therapy. Materials and Methods Reagents and Antibodies. MEM and FCS were obtained from Life Technologies, Inc., Grand Island, NY. hrigfbp-3 was from Celtrix Laboratories, Richmond, VA. Human recombinant IGF-I and IGF-II were from GroPep, Adelaide, Australia. Rabbit anti-phospho Akt (Ser473), mouse antiphospho p44/42 MAPK (Thr202/Tyr204), rabbit anti-akt, and mouse anti-mapk antibodies were purchased from New England Biolabs Inc., Beverly, MA. Rabbit anti-phospho Elk-1 (Ser383), rabbit antihuman IGF-I, antihuman IGF-IR, mouse anti- -tubulin, and protein A-agarose were obtained from Santa Cruz Biotechnology, Santa Cruz, CA. Rabbit antihuman IGFBP-3, rabbit anti-irs-1 and mouse antiphosphotyrosine 4G10 were from Upstate Biotechnology, Lake Placid, NY. Rabbit antihuman IGF-II were from Austral Biologicals, San Ramon, CA. All of the antibodies were used at an indicated final concentration. Horseradish peroxidaseconjugated donkey antimouse or antirabbit secondary antibodies were from Pierce, Rockford, IL. Chemiluminescent detection system was from Amersham, Pharmacia Biotech,

7 Cell Growth & Differentiation 121 Arlington Heights, IL. L-phosphatidyl inositol-4 monophosphate was from Sigma Chemical Co., St. Louis, MO. Collection of Human HCC and Adjacent Nontumor Liver Specimens. Tissue samples were obtained intraoperatively from tumors and adjacent nontumor livers during liver resection for HCC in 28 patients at the Singapore General Hospital. The samples were snap frozen in liquid nitrogen and stored at 80 C until analysis. A similar set of samples was fixed in 10% formalin and paraffin embedded. The diagnosis of HCC was confirmed histologically in all of the cases. Prior written informed consent was obtained from all of the patients, and the study received ethics board approval at both institutions. Immunolocalization of IGFBP-3. Formalin- and paraffinembedded sections were used for IGFBP-3, IGF-II, and IGF-I immunolocalization. This was performed using rabbit antihuman IGFBP-3 (1:500 dilution), rabbit antihuman IGF-I (3 g/ ml), and rabbit antihuman IGF-II (3 g/ml) antibodies as described (28). Nonspecific staining was evaluated for each specimen using either a similar concentration of IgG or by absorbing the primary with appropriate specific immunogen. The slides were evaluated, and intensity of the staining was scored. Specific staining was semiquantitated by assigning a score of 0 to based on increasing green fluorescence intensity. The results shown in Table 1 represent the average score obtained from twice staining. Cell Culture. Human hepatoma HepG2 cells were obtained from American Type Culture Collection and maintained as monolayer cultures in MEM supplemented with 10% FCS growth medium. For proliferation study, confluent cultures of HepG2 cells were trypsinized and plated at cells in 24-well plates (Nunc, Nalgene Nunc International, Rochester, NY) with growth medium. After 48 h, the cell monolayers were rinsed twice with MEM PSF medium and incubated additionally in PSF medium for 24 h. After 24 h, various concentrations of hrigfbp-3, human recombinant IGF-I, or 25 ng/ml human recombinant IGF-I in conjunction with various concentrations of hrigfbp-3 were added in triplicate in specific pathogen-free medium for 48 h. Cell number was determined as described previously (28). To determine the effect of exogenous IGF-II and autocrine IGF-II production on HepG2 cell proliferation, HepG2 cells were plated and grown as described above, and the cells subsequently treated with rabbit preimmune serum (1:800 dilution), 25 ng/ml human recombinant IGF-II, rabbit antihuman IGF-II antibody (2 g/ml), or with combination for 48 h. Cell number was determined as described above. Western Blotting. To determine changes in the expression of IGFBP-3 and IGF-IR, snap-frozen HCC tumors and normal adjacent liver tissues were thawed and homogenized in lysis buffer (1 mm CaCl 2,1mM MgCl 2, 1% NP40, 1 g/ml leupeptin, 1 g/ml aprotinin, 1 M phenylmethylsulfonyl fluoride, and 100 M NaVO 4 ). Proteins were subjected to Western blot analysis as described (28). Blots were incubated with 1:5000 antihuman IGFBP-3, 1 g/ml antihuman IGF-IR, 0.5 g/ml anti- -tubulin antibodies, and 1:7500 horseradish peroxidase-conjugated donkey antimouse or antirabbit secondary antibody. The blots were then visualized with a chemiluminescent detection system as described by the manufacturer. To examine the effects of IGF-I and IGFBP-3 on IGF-IR, IRS-1, Elk-1, Akt-1, MAPK, and their phosphorylation forms, HepG2 cells were plated at cells per Petri dishes (Becton Dickinson, Lincoln Park, NJ) in growth medium. After 48 h, the cell monolayers were rinsed twice with MEM SFM and additionally incubated in PSF medium for 24 h. After 24 h, cells were treated with 250 ng/ml hrigfbp-3 in SFM for 24 h. Cells were rinsed once and then incubated with 25 ng/ml IGF-I for 10 min. After IGF-I stimulation, cells were harvested and lysed in lysis buffer as described above. To determine the effects of IGFBP-3 on IGF-I-induced MAPK, Elk-1, and Akt phosphorylation, Western blotting was performed using phospho Akt (Ser473; 1 g/ml), phospho p44/42 MAPK (Thr202/Tyr204; 1 g/ml) and phospho Elk-1 (Ser383; 1 g/ml) antibodies. Levels of IGF-IR subunit and IRS-1, and their phosphorylated forms were determined by immunoprecipitation of total cell lysates using anti-igf-ir (2 g/ml) and IRS-1 (4 g/ml) antibodies, respectively, as described by the supplier. Immunoprecipitated proteins was blotted using rabbit antihuman IGF-IR (1 g/ml), antihuman IRS-1 (1 g/ml), and antiphosphotyrosine 4G10 antibodies (1 g/ml). To study IRS/PI3k p85 interactions, cell lysates were precipitated with anti-irs-1 antibody. Immunocomplexes were used to determined PI3k activity associated with IRS-1. To measure PI3k activity, 500 g of total protein were incubated with 4 g of anti-irs-1 antibody for 3 h. The antigen/antibody complex was precipitated with 50 l of protein A- agarose and washed with kinase buffer [25 mm Tris (ph 7.5), 5mM -glycerol phosphate, 2 mm DTT and 0.1 mm Na 3 V0 4 ]. Immunoprecipitates were resuspended in 25 l of kinase buffer containing 20 g of L-phosphatidyl inositol-4 monophosphate and 5 l of 100 mm MgCl 2, and then adding 30 Ci of [ 32 P]ATP in 2.5 l of 0.88 mm ATP per kinase reaction. Kinase products were resolved using TLC with a CHCL 3 /methanol/h 2 O/NH 4 OH (40:48:10:5) solvent system. Quantitative analysis of 32 P-labeled inositol 1,3,4 5 phosphate levels was determined by scanning of the blots densitometrically. Statistical Analysis. For quantitation analysis, the sum of the density of bands corresponding to protein blotting with the antibody under study was calculated and normalized the amount of -tubulin. Differences in cell number and the levels of proteins under studied were analyzed by the Mann- Whitney U test. References 1. Okuda, K., Ohtsuki, T., Obata, H., Tomimatsu, M., Okazaki, N., Hasegawa, H., Nakajima, Y., and Ohnishi, K. Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Study of 850 patients. Cancer (Phila.), 56: , Lai, E. C., Fan, S. T., Lo, C. M., Chu, K. M., Liu, C. L., and Wong, J. Hepatic resection for hepatocellular carcinoma. An audit of 343 patients. Ann. Surg., 221: , Takenaka, K., Kawahara, N., Yamamoto, K., Kajiyama, K., Maeda, T., Itasaka, H., Shirabe, K., Nishizaki, T., Yanaga, K., and Sugimachi, K. Results of 280 liver resections for hepatocellular carcinoma. Arch. Surg., 131: 71 76, Huguet, C., Stipa, F., and Gavelli, A. Primary hepatocellular cancer: Western experience. In: L. Blumgart (ed.), Surgery of the Liver and Biliary Tract, pp Churchill Livingstone: London, 2000.

8 122 IGFBP-3 in Hepatocellular Carcinoma 5. Lai, E., and Wong, J. Hepatocellular carcinoma: the Asian experience. In: L. Blumgart (ed.), Surgery of the Liver and the Biliary Tract, pp London: Churchill Livingstone, Chan, E. S. Y., Chow, P. K. H., Tai, B. C., Martin, D., and Soo, K. C. Neoadjuvant and adjuvant therapy for operable hepatocellular carcinoma. (Cochrane Review). In: The Cochrane Library, 1, Oxford. 7. Scharf, J. G., Schmidt-Sandte, W., Pahernik, S. A., Ramadori, G., Braulke, T., and Hartmann, H. Characterization of the insulin-like growth factor axis in a human hepatoma cell line (PLC). Carcinogenesis (Lond.), 19: , Tsai, T. F., Yauk, Y. K., Chou, C. K., Ting, L. P., Chang, C., Hu, C. P., Han, S. H., and Su, T. S. Evidence of autocrine regulation in human hepatoma cell lines. Biochem. Biophys. Res. Commun., 153: 39 45, Su, T. S., Liu, W. Y., Han, S. H., Jansen, M., Yang-Fen, T. L., P eng, F. K., and Chou, C. K. Transcripts of the insulin-like growth factors I and II in human hepatoma. Cancer Res., 49: , Ng, I. O., Lee, J. M., Srivastava, G., and Ng, M. Expression of insulinlike growth factor II mrna in hepatocellular carcinoma. J. Gastroenterol. Hepatol., 13: , Cariani, E., Lasserre, C., Seurin, D., Hamelin, B., Kemeny, F., Franco, D., Czech, M. P., Ullrich, A., and Brechot, C. Differential expression of insulin-like growth factor II mrna in human primary liver cancers, benign liver tumors, and liver cirrhosis. Cancer Res., 48: , Rogler, C. E., Yang, D., Rossetti, L., Donohoe, J., Alt, E., Chang, C. J., Rosenfeld, R., Neely, K., and Hintz, R. Altered body composition and increased frequency of diverse malignancies in insulin-like growth factor-ii transgenic mice. J. Biol. Chem., 269: , Khandwala, H. M., McCutcheon, I. E., Flyvbjerg, A., and Friend, K. E. The effects of insulin-like growth factors on tumorigenesis and neoplastic growth. Endocr. Rev., 21: , Clemmons, D. R. Insulin-like growth factor binding proteins and their role in controlling IGF actions. Cytokine Growth Factor Rev., 8: 45 62, Arany, E., Afford, S., Strain, A. J., Winwood, P. J., Arthur, M. J., and Hill, D. J. Differential cellular synthesis of insulin-like growth factor binding protein-1 (IGFBP-1) and IGFBP-3 within human liver. J. Clin. Endocrinol. Metab, 79: , Huynh, H., Nickerson, T., Yang, X., and Pollak, M. Regulation of insulin-like growth factor I receptor by the pure antiestrogen ICI 182,780. Clin. Cancer Res., 2: , Huynh, H., Yang, X., and Pollak, M. Estradiol and antiestrogens regulate a growth inhibitory insulin- like growth factor binding protein 3 autocrine loop in human breast cancer cells. J. Biol. Chem., 271: , Oh, Y., Mullers, H. L., Ng, L., and Rosenfeld, R. G. Transforming growth factor-b-induced cell growth inhibition in human breast cancer cells is mediated through insulin-like growth factor-binding protein-3 action. J. Biol. Chem., 270: , LeRoith, D., Adamo, M. L., Shemer, J., Lanau, F., Shen-Orr, Z., Yaron, A., Roberts, C. T., Clemmons, D. R., Sheikh, M. S., and Shao, Z. M. Retinoic acid inhibits growth of breast cancer cell lines: the role of insulinlike growth factor binding proteins. Growth Regul., 3: 78 80, Herman, J. G., Latif, F., Weng, Y., Lerman, M. I., Zbar, B., Liu, S., Samid, D., Duan, D. S., Gnarra, J. R., Linehan, W. M. et al. Silencing of the VHL tumor-suppressor gene by DNA methylation in renal carcinoma1070. Proc. Natl. Acad. Sci. USA, 91: , Huynh, H., Alpert, L., and Pollak, M. Silencing of the mammary derived growth inhibitor (MDGI) gene in breast neoplasms is associated with epigenetic changes. Cancer Res., 56: , Wong, I. H., Johnson, P. J., Lai, P. B., Lau, W. Y., and Lo, Y. M. Tumor-derived epigenetic changes in the plasma and serum of liver cancer patients. Implications for cancer detection and monitoring. Ann N. Y. Acad Sci., 906: , Kamb, A., Gruis, N. A., Weaver-Feldhaus, J., Liu, Q., Harshman, K., Tavtigian, S. V., Stockert, E., Day, R. S., III, Johnson, B. E., and Skolnick, M. H. A cell cycle regulator potentially involved in genesis of many tumor types. Science (Wash. DC), 264: , Montesano, R., Hainaut, P., and Wild, C. P. Hepatocellular carcinoma: from gene to public health. J. Natl. Cancer Inst., 89: , Chen, C. J., Yu, M. W., and Liaw, Y. F. Epidemiological characteristics and risk factors of hepatocellular carcinoma. J. Gastroenterol. Hepatol., 12: S294 S308, Oh, Y., Muller, H. L., Lamson, G., and Rosenfeld, R. G. Insulin-like growth factor (IGF)-independent action of IGF-binding protein-3 in Hs578T human breast cancer cells. J. Biol. Chem., 268: , Oh, Y., Muller, H. L., Pham, H., and Rosenfeld, R. G. Demonstration of receptors for insulin-like growth factor binding protein-3 on Hs578T human breast cancer cells. J. Biol. Chem., 268: , Huynh, H., Larsson, C., Narod, S., and Pollak, M. Tumor suppressor activity of the gene encoding mammary-derived growth inhibitor. Cancer Res., 55: , 1995.